40% Unemployment or a 3-Day Week? It's the Same Maths

40% unemployment and a 3-day work week are the same thing. That is the argument from economist Alex Tabarrok in Fortune this week. Same total hours of work. Same economic output. The only difference is who gets the time. It is a genuinely useful way to think about what is happening. And I suspect it is landing badly for a lot of people.

The concern is real

If you lead a team or run an organisation, this framing probably does not feel reassuring. It feels like someone telling you that catastrophe and paradise are a coin flip. According to Access Group research, 41% of UK employees already fear AI's impact on their jobs, while only 25% of HR leaders share that concern. That gap matters. People are worried, and they are picking up on signals their leaders are not sending. When you read that 40% of the workforce could be unemployed, the maths being elegant does not make the fear smaller. It makes it sharper.

Here's what I think is actually going on

Tabarrok's insight is genuinely important, but I think the framing misses something. He presents unemployment and shorter hours as two endpoints of the same equation. That is mathematically true. But it suggests the outcome is binary - catastrophe or wonderland - when actually it is a spectrum of thousands of small decisions made by individual leaders in individual organisations every week.

Here is the bit that matters. Between 1870 and today, working hours fell by roughly 40% - from nearly 3,000 hours per year to about 1,800. Unemployment did not rise to match. We did not get 40% unemployment. We got weekends. We got holidays. We got the idea that working five days instead of six was normal. That transition happened because of choices - by unions, by governments, by employers who realised that rested workers were better workers. The same kind of choices are available now. The question is whether leaders make them deliberately or let the default setting - which is always redundancy - play out.

Will AI cause mass unemployment?

Not if leaders treat AI productivity gains the way previous generations treated mechanisation - by redistributing time rather than eliminating people. The maths that produces 40% unemployment is identical to the maths that produces a 3-day week. The variable is not technology or economics. It is leadership. Organisations that invest in AI training for their people and redesign roles around the time AI frees up will keep their best talent and find new capacity they never had.

The path forward

There are a couple of ways to think about this practically.

First, look at what AI is actually compressing in your organisation. Not theoretically - right now. Where are people spending hours on work that AI could reduce to minutes? That is your freed capacity. What you do with it is the decision that determines whether your people experience this as a threat or an opportunity.

Second, close the training gap. Only 19% of employees have received formal AI training, according to Access Group. That means the vast majority of your workforce is watching AI arrive without any framework for understanding what it means for them. Fear fills that vacuum fast. Giving people practical capability is the single most effective way to turn anxiety into agency.

The UK's four-day week trials already showed this pattern in miniature. Revenue stayed broadly flat while staff turnover dropped by 57%. When you give people time back and trust them with it, they do not become less productive. They become more committed.

What this comes down to

The economist is right about the maths. Forty percent unemployment and a 3-day week really are the same equation. But equations do not make choices. Leaders do. The shift from human doing to human being is not something that happens to us. It is something we design. If you are a senior leader working through what AI means for your organisation, the AI Leaders Fellowship is where we work through exactly these decisions - practically, not theoretically.